融合传统与前馈方法，提升3D重建在复杂场景下的精度与鲁棒性。
Global Structure-from-Motion Meets Feedforward Reconstruction
- 结合经典SfM与前馈重建优势，构建新流水线。
- 在低纹理、重叠少等挑战场景下表现最优，跨数据集领先。
- 适合需要高精度3D重建的科研与工业应用。
从图像中同时估计相机位姿和三维场景结构的运动恢复结构（SfM）仍是计算机视觉的核心挑战，存在诸多未解问题。近期前馈3D重建方法在低纹理、重叠度低及对称性等经典SfM易失败场景中取得显著进展。然而，前馈方法通常在可扩展性、准确率或鲁棒性上受限，且在标准重建场景中仍不及传统方法。本文系统分析其局限性，提出一种融合经典与前馈方法优势的新SfM流程。在多个数据集上的大量实验表明，该方法在广泛场景中达到当前最优性能。代码已开源：https://github.com/colmap/gluemap。
原文摘要 · Abstract (English)
Structure-from-Motion -- the process of simultaneously estimating camera poses and 3D scene structure from a collection of images -- remains a central challenge in computer vision, with many open problems yet to be solved. Recent advances in feedforward 3D reconstruction have made significant strides in overcoming persistent failure cases of classical SfM methods, particularly in scenarios characterized by low texture, limited overlap, and symmetries. However, while feedforward approaches excel in these challenging conditions, they often face limitations regarding scalability, accuracy, or robustness, and typically fall short of classical methods in standard reconstruction settings. In this work, we systematically analyze these limitations and propose a new Structure-from-Motion pipeline by combining the respective strengths of classical and feedforward methods. Extensive experiments across multiple datasets show the benefits of our approach, achieving state-of-the-art results across a wide range of scenarios. We share our system as an open-source implementation at https://github.com/colmap/gluemap.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。